Prepare and test your skills

Prepare and test your skills

Differentiate Management Responsibilities and Control means understanding how work is divided between the cloud provider and the customer in each cloud model. The three models are Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). They form a spectrum where customer control decreases from IaaS to SaaS while ease of use increases, and administrative work shifts from the customer to the provider.
Infrastructure as a Service (IaaS) is the cloud model that gives the customer the greatest configuration control. The provider owns the physical hardware, networking, and virtualization, while the customer manages the operating systems, applications, data, storage configuration, and database connections that run on top. Provisioning IaaS often means creating virtual machines and networks, so the customer is also responsible for patching the operating system and configuring the application stack. Tools like Terraform and Infrastructure Manager (opens in a new tab) help manage IaaS at scale by defining the desired end state in configuration files called blueprints. Infrastructure Manager deploys those blueprints through deployments and revisions, but it manages only the lifecycle of infrastructure resources and does not deploy applications onto them.
Platform as a Service (PaaS) is the cloud model where the provider manages hardware, operating systems, and runtimes so the customer can focus on code and data. The customer deploys application code or containers to the platform, and traffic flows from users to the provider-managed platform to the customer code. The customer no longer patches servers, manages the runtime, or provisions capacity for the platform itself. Google Cloud examples show this boundary clearly: App Engine is a fully managed serverless platform for building highly scalable applications, Cloud Run is a fully managed serverless platform for developing and deploying highly scalable containerized applications, and Cloud Run functions is a functions-as-a-service offering that runs event-driven code with zero server management on a pay-as-you-go model.
Software as a Service (SaaS) is the cloud model that prioritizes ease of use with the lowest administrative overhead. The provider manages almost everything, including hardware, operating systems, application code, updates, and storage, while the customer only configures settings and supplies data. The customer never manages servers, runtimes, or application logic. Terraform Cloud is a SaaS application that runs Terraform in a stable remote environment and securely stores state and secrets. Data flows from a connected version control system such as GitHub, GitLab, or Bitbucket to a Terraform Cloud workspace, where new commits can automatically trigger Terraform plans, and the service also integrates with the Terraform CLI and an API.
How control and effort compare across models depends on who manages hardware, operating systems, and application code. IaaS keeps operating systems and applications with the customer, PaaS keeps only code and data with the customer, and SaaS keeps only usage and settings with the customer. The following comparison summarizes the tradeoff:
| Model | Customer manages | Provider manages | Control versus ease |
|---|---|---|---|
| Infrastructure as a Service (IaaS) | Operating systems, applications, data, and infrastructure configuration | Physical hardware, networking, and virtualization | Greatest configuration control, highest overhead |
| Platform as a Service (PaaS) | Application code, data, and application settings | Hardware, operating systems, runtimes, and scaling | Balanced control, reduced overhead |
| Software as a Service (SaaS) | User settings, data input, and access configuration | Hardware, operating systems, application code, updates, and state storage | Least control, lowest overhead |
Choose IaaS when the workload needs custom operating systems or deep infrastructure control. Choose PaaS when the team wants to deliver code without managing servers, and choose SaaS when the goal is to consume complete software without operating any part of it.
Exam tip: IaaS leaves operating systems and applications with the customer, PaaS leaves only code and data with the customer, and SaaS leaves only usage and settings with the customer.
Assess Scalability and Operational Trade-offs means comparing how each cloud model handles growth and daily upkeep. Scalability is the ability to handle more users, data, or traffic, while operational trade-offs describe who performs the scaling and maintenance. The central tradeoff is flexibility and control of managing your own environment versus the efficiency of consuming managed services.
Manual scaling and maintenance in IaaS reflects that the customer controls virtual machines, storage, networking, operating systems, and applications. Scaling often requires manual steps such as adding or removing virtual machines and adjusting load balancing as demand changes. Maintenance tasks such as patching, updates, and monitoring also remain with the customer. This manual work is time consuming and error prone at scale, so choose IaaS when deep control over the environment justifies the operational investment.
Automated scaling and management in PaaS reflects that the provider handles the underlying infrastructure, operating system, and runtime. The customer focuses on deploying and managing applications while the platform adds or removes resources as demand changes and maintains the platform itself. Infrastructure Manager (opens in a new tab) shows this automation: it validates deployment requests, creates a Cloud Build job, runs Terraform init, validate, and apply, streams logs, and retries certain failures automatically. This removes manual steps compared with managing the same deployment by hand. Choose PaaS when automation and efficiency matter but the team still needs to configure applications and some resources.
Fully managed operations in SaaS means the provider handles scaling, maintenance, and updates for a complete application. The customer simply uses the software through a browser or API and does not manage infrastructure or platforms. Terraform Cloud is a SaaS application that runs Terraform remotely, securely stores state and secrets, integrates with version control systems, and offers an API for existing workflows. Because the provider handles all operations, SaaS offers the highest efficiency and lowest burden but the least flexibility over the environment. Choose SaaS when a ready-made service meets the need and minimizing operations is the priority.
Automation maturity from Adopt to Scale describes how Infrastructure as Code (IaC) (opens in a new tab) practices reduce manual work over time. Infrastructure as Code provisions and manages infrastructure with versioned, reusable configuration files instead of manual console or script steps. The Terraform on Google Cloud maturity model (opens in a new tab) defines three stages that increase automation and consistency:
Moving from Adopt to Scale requires investment in tools and processes, but it improves consistency and reduces error-prone manual work as the number of resources and teams grows.
Balancing redundancy and efficiency means weighing readiness against wasted resources. Resource redundancy is a primary driver of energy waste when idle capacity is kept running. An active-passive setup keeps resources in the passive site idle, which wastes energy, while an optimally sized active-active architecture lets provisioned resources across regions actively serve traffic. Teams should replicate data and services across regions only when needed for high availability or disaster recovery. For disaster recovery, IaC can rapidly provision environments only when needed instead of keeping a warm site running.
Exam tip: IaaS gives the most control but requires manual scaling and maintenance; PaaS and SaaS shift scaling and maintenance to the provider, trading control for efficiency.
Evaluate Business Use Cases for Model Selection means matching a workload and business goal to IaaS, PaaS, or SaaS. Each model places the shared responsibility boundary in a different place, so the choice depends on how much control the workload needs, what skills the team has, and how quickly results are needed. Moving from IaaS to PaaS to SaaS reduces operational work but also reduces direct control.
Where the shared responsibility boundary sits determines which layers the customer must staff, patch, and monitor. In IaaS, the provider manages physical data centers, networking, storage hardware, and virtualization, while the customer manages the operating system, middleware, runtime, applications, and data. In PaaS, the provider manages everything up through runtime and middleware, so the customer manages mainly application code, configuration, and data. In SaaS, the provider manages the entire stack including the application, and the customer manages only its own data, user accounts, and access settings.
Migrating legacy workloads with IaaS fits applications that depend on a specific operating system version, custom middleware, or server-level settings. A lift-and-shift migration moves the application to a cloud virtual machine with minimal changes, which avoids rewriting code and keeps migration fast. The tradeoff is that operational work remains: the organization still patches the operating system, manages scaling, and handles backups for that server. Choose this path when preserving existing behavior matters more than reducing server management.
Building cloud-native applications with PaaS fits new development where speed-to-market matters. Developers write and deploy code while the platform supplies the runtime, scaling, and patching underneath, so the team spends time on features instead of infrastructure. Traffic flows from users to the managed platform to the new application code without the team provisioning servers. The tradeoff is less control over the runtime environment, so applications with unusual platform requirements may not fit.
Deploying productivity tools with SaaS fits standard software such as email, document editing, and collaboration tools. The software already exists and the provider runs it, so users consume it directly with no servers, operating systems, or application code for the organization to manage. This is the fastest path to value and the lightest demand on internal staff, which matters for teams with limited technical resources. The tradeoff is the least control, because the organization accepts the provider's features, update schedule, and configuration options.
A four-step sequence for choosing a model applies control, skills, speed, and fit in order. It connects organizational goals like speed-to-market and resource constraints to the technical boundary:
A decision tree branching on control needs, staff skills, delivery speed, and PaaS runtime fit to select IaaS for legacy lift-and-shift, PaaS for new code, or SaaS for ready-made tools.
Cost also shifts with the model because IaaS charges for infrastructure the customer still operates, while PaaS and SaaS bundle more management into the price.
Exam tip: Lift-and-shift of an existing server-based application fits IaaS, a newly built application that should not require server management fits PaaS, and ready-made software consumed directly by users fits SaaS.
Infrastructure as a Service (IaaS) gives the customer the greatest configuration control because the customer manages operating systems, applications, data, and infrastructure configuration while the provider manages only physical hardware, networking, and virtualization. Platform as a Service (PaaS) shifts more work to the provider, which manages hardware, operating systems, runtimes, and scaling while the customer manages only application code, data, and application settings. Software as a Service (SaaS) has the lowest administrative overhead because the provider manages almost everything including application code, updates, and storage while the customer only configures settings and supplies data.
Choose Infrastructure as a Service (IaaS) when the workload needs custom operating systems, middleware, or deep infrastructure control, such as a lift-and-shift migration of a legacy application that depends on a specific operating system version or server-level settings. Choose Platform as a Service (PaaS) for new cloud-native development where the team wants to deliver code without managing servers, and choose Software as a Service (SaaS) for standard productivity tools such as email, document editing, and collaboration that users can consume directly.
Infrastructure as a Service (IaaS) typically requires manual scaling and maintenance because the customer controls virtual machines, operating systems, and applications and must add or remove machines, adjust load balancing, and perform patching and monitoring. Platform as a Service (PaaS) automates scaling and platform maintenance because the provider handles the underlying infrastructure, operating system, and runtime while the customer focuses on deploying applications. Software as a Service (SaaS) leaves all scaling, maintenance, and updates to the provider because the customer simply uses the complete application through a browser or API.
Last updated